[Gate-users] Sensitivity discrepancies in optical vs non-optical simulations

[Devin VanElburg]

Hello all,


I've been checking the relative sensitivity of a PET geometry by taking the ratio of true events to total events. I checked this with both optical photons and non-optical photon simulations and got significantly different results. I am using LYSO as the scintillator and a small acrylic cube as a phantom. The prevalence of Compton events in the phantom using the attachPhantomSD was negligible compared to the crystal for both the optical and non-optical simulations. I used a Na-22 source at the centre of the FOV. 


With optical processes on, I got a sensitivity of about 90% (so 10% Compton) for 100,000 coincident events. However without optical processes on, I got a sensitivity of about 35% (65% Compton) for 1mil events. 


The only differences in the macro files are defining all the surfaces in the optical files and the other unique optical commands. The geometries, physics lists, and digitizers (minus opticaladder of course) were identical.


Where could this huge discrepancy be coming from? I appreciate any insight that anyone can give me.


Cheers,
Devin Van Elburg

------------------------
Physics w/ c/ Biomedical Physics
Lakehead University


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opengatecollaboration.org/mailman/private/gate-users/attachments/20160216/e7b5ebb3/attachment.html>